[heap] Switch to 500k pages

[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org

Committed: https://crrev.com/ffe5c670e1559d11e7b252e15fec38765e7dbe4f
Cr-Commit-Position: refs/heads/master@{#38533}

[Michael Lippautz, 2016-05-31 09:19:34]

Description was changed from

==========
[heap] Switch to 500k pages

BUG=
==========

to

==========
[heap] Switch to 500k pages

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-06-29 13:39:56]

Patchset #10 (id:180001) has been deleted

[Michael Lippautz, 2016-07-29 11:52:53]

Description was changed from

==========
[heap] Switch to 500k pages

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k.

Perf sheriffs: This potentially tanks (single-digit %) time-based performace
benchmarks, at the benefit of decreasing fragmentaton on V8's heap.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-07-29 11:54:19]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k.

Perf sheriffs: This potentially tanks (single-digit %) time-based performace
benchmarks, at the benefit of decreasing fragmentaton on V8's heap.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Perf sheriffs: This potentially tanks (expected: small single-digit %)
time-based performace benchmarks, at the benefit of decreasing fragmentaton on
V8's heap. Please open separate bugs for upcoming issues.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-07-29 11:55:10]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Perf sheriffs: This potentially tanks (expected: small single-digit %)
time-based performace benchmarks, at the benefit of decreasing fragmentaton on
V8's heap. Please open separate bugs for upcoming issues.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Furthermore, this also enables us to get rid of the new space border page (age
mark) by keeping the same heap size

Perf sheriffs: This potentially tanks (expected: small single-digit %)
time-based performace benchmarks, at the benefit of decreasing fragmentaton on
V8's heap. Please open separate bugs for upcoming issues.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-07-29 12:07:27]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Furthermore, this also enables us to get rid of the new space border page (age
mark) by keeping the same heap size

Perf sheriffs: This potentially tanks (expected: small single-digit %)
time-based performace benchmarks, at the benefit of decreasing fragmentaton on
V8's heap. Please open separate bugs for upcoming issues.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Furthermore, this also enables us to get rid of the new space border page (age
mark) by keeping the same heap size.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-08-01 10:46:28]

Patchset #10 (id:200001) has been deleted

[Michael Lippautz, 2016-08-01 10:46:37]

Patchset #9 (id:160001) has been deleted

[Michael Lippautz, 2016-08-01 10:46:46]

Patchset #7 (id:120001) has been deleted

[Michael Lippautz, 2016-08-01 10:46:53]

Patchset #7 (id:140001) has been deleted

[Michael Lippautz, 2016-08-02 10:48:42]

Patchset #14 (id:360001) has been deleted

[Michael Lippautz, 2016-08-02 10:49:57]

Patchset #2 (id:20001) has been deleted

[Michael Lippautz, 2016-08-02 11:53:37]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

Furthermore, this also enables us to get rid of the new space border page (age
mark) by keeping the same heap size.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page and still keep
the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-08-02 11:53:56]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page and still keep
the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page and still
 keep the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-08-02 11:54:11]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page and still
 keep the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

[Michael Lippautz, 2016-08-02 11:56:13]

Patchset #8 (id:260001) has been deleted

[Michael Lippautz, 2016-08-02 11:56:29]

Patchset #8 (id:280001) has been deleted

[Michael Lippautz, 2016-08-02 11:56:37]

Patchset #8 (id:300001) has been deleted

[Michael Lippautz, 2016-08-02 11:56:59]

Patchset #5 (id:100001) has been deleted

[Michael Lippautz, 2016-08-02 12:03:17]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

BUG=chromium:581412
LOG=N
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Results in a 5-10% improvement on v8.infinite_scroll evacuation bucket.

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

[Michael Lippautz, 2016-08-02 12:03:17]

mlippautz@chromium.org changed reviewers:
+ hpayer@chromium.org, ulan@chromium.org

[Michael Lippautz, 2016-08-02 12:03:30]

ptal (still fixing tests)

[Michael Lippautz, 2016-08-02 12:41:44]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-02 12:41:54]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-02 13:40:11]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-02 13:40:12]

Dry run: Try jobs failed on following builders:
  v8_win_nosnap_shared_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)
  v8_win_nosnap_shared_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)

[Michael Lippautz, 2016-08-02 19:19:45]

Patchset #5 (id:220001) has been deleted

[Michael Lippautz, 2016-08-03 07:41:20]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Results in a 5-10% improvement on v8.infinite_scroll evacuation bucket.

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

[Michael Lippautz, 2016-08-03 07:41:31]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

[Michael Lippautz, 2016-08-03 10:26:35]

It's ready now.

[ulan, 2016-08-03 10:31:51]

lgtm

https://codereview.chromium.org/2013713003/diff/440001/src/heap/spaces.cc
File src/heap/spaces.cc (right):

https://codereview.chromium.org/2013713003/diff/440001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1231: if (needed / area_size > 0) return area_size;
if (need >= area_size) return area_size

[Michael Lippautz, 2016-08-03 12:44:43]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-03 12:44:51]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-03 12:58:22]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-03 12:58:23]

Dry run: Try jobs failed on following builders:
  v8_linux64_avx2_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_avx2_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng_trig...)

[Michael Lippautz, 2016-08-04 08:43:57]

https://codereview.chromium.org/2013713003/diff/440001/src/heap/spaces.cc
File src/heap/spaces.cc (right):

https://codereview.chromium.org/2013713003/diff/440001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1231: if (needed / area_size > 0) return area_size;
On 2016/08/03 10:31:51, ulan wrote:
> if (need >= area_size) return area_size

Done.

[Michael Lippautz, 2016-08-04 13:08:58]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-04 13:09:08]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-04 13:17:23]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-04 13:17:24]

Dry run: Try jobs failed on following builders:
  v8_win_compile_dbg on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_compile_dbg/builds/2...)

[Michael Lippautz, 2016-08-04 17:38:24]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-04 17:38:34]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-04 18:14:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-04 18:14:25]

Dry run: Try jobs failed on following builders:
  v8_win_nosnap_shared_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)
  v8_win_nosnap_shared_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)

[Michael Lippautz, 2016-08-08 15:25:48]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-08 15:25:55]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-08 16:24:33]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-08 16:24:34]

Dry run: This issue passed the CQ dry run.

[Hannes Payer (out of office), 2016-08-08 16:25:03]

Cool stuff! Some comments.

https://codereview.chromium.org/2013713003/diff/520001/src/base/build_config.h
File src/base/build_config.h (right):

https://codereview.chromium.org/2013713003/diff/520001/src/base/build_config....
src/base/build_config.h:203: const int kPageSizeBits = 19;
Would it be possible to make the page size configurable, with a flag, to
experiment... ?

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc
File src/heap/spaces.cc (right):

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1223: if (snapshot_requires > 0) {
I think this if case should be a DCHECK and the else default should not exist,
i.e. the function should only be called if we are deserializing a snapshot.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1227: memory_served += page->area_size();
size of actual objects would be more precise here.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1230: const int area_size =
MemoryAllocator::PageAreaSize(identity());
Just call AreaSize() here.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1236: return static_cast<int>(needed);
Maybe we should do a different approach here:
How about just allocated as much memory as needed and then shrink the pages
after deserialization?

[Michael Lippautz, 2016-08-08 18:22:46]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-08 18:22:59]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-08 18:54:15]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-08 18:54:16]

Dry run: Try jobs failed on following builders:
  v8_linux_arm_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng/builds/...)
  v8_linux_arm_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng_trigger...)

[Michael Lippautz, 2016-08-09 08:00:34]

Patchset #17 (id:560001) has been deleted

[Michael Lippautz, 2016-08-09 08:04:15]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-09 08:04:27]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-09 08:22:16]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-09 08:22:17]

Dry run: Try jobs failed on following builders:
  v8_linux64_avx2_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_avx2_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng_trig...)

[Michael Lippautz, 2016-08-09 11:00:57]

Patchset #19 (id:620001) has been deleted

[Michael Lippautz, 2016-08-09 11:30:24]

mlippautz@chromium.org changed reviewers:
+ yangguo@chromium.org

[Michael Lippautz, 2016-08-09 11:30:26]

+Yang

PTAL: We now shrink the pages after deserialization the isolate snapshot. I
would like to land this and see what the actual issues are (e.g. page cycler)
retrospectively. The subtests I looked at in page_cycler.top_25 all have
>15-20MB used heap.

If it does not work out, we can always leave more room on the last page after
deserialization.

https://codereview.chromium.org/2013713003/diff/520001/src/base/build_config.h
File src/base/build_config.h (right):

https://codereview.chromium.org/2013713003/diff/520001/src/base/build_config....
src/base/build_config.h:203: const int kPageSizeBits = 19;
On 2016/08/08 16:25:02, Hannes Payer wrote:
> Would it be possible to make the page size configurable, with a flag, to
> experiment... ?

There's the issue with kMaxRegularHeapObjectSize and the constant in objects.h
that is only related via comment because we don't want to include the header. 

Having that said, after this CL, changing the page size should merely be a
3-line CL (3 constants).

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc
File src/heap/spaces.cc (right):

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1223: if (snapshot_requires > 0) {
On 2016/08/08 16:25:02, Hannes Payer wrote:
> I think this if case should be a DCHECK and the else default should not exist,
> i.e. the function should only be called if we are deserializing a snapshot.

Obsolete.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1227: memory_served += page->area_size();
On 2016/08/08 16:25:02, Hannes Payer wrote:
> size of actual objects would be more precise here.

Obsolete.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1230: const int area_size =
MemoryAllocator::PageAreaSize(identity());
On 2016/08/08 16:25:02, Hannes Payer wrote:
> Just call AreaSize() here.

Obsolete.

https://codereview.chromium.org/2013713003/diff/520001/src/heap/spaces.cc#new...
src/heap/spaces.cc:1236: return static_cast<int>(needed);
On 2016/08/08 16:25:02, Hannes Payer wrote:
> Maybe we should do a different approach here:
> How about just allocated as much memory as needed and then shrink the pages
> after deserialization?

Done.

[Michael Lippautz, 2016-08-09 12:35:07]

Patchset #13 (id:480001) has been deleted

[Yang, 2016-08-09 13:26:29]

some nits. otherwise lgtm.

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot.h
File src/snapshot/snapshot.h (right):

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
src/snapshot/snapshot.h:88: // [6] number of contexts N
Please update the indices here.

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
src/snapshot/snapshot.h:97: static const int kNumPagedSpaces = LAST_PAGED_SPACE
- FIRST_PAGED_SPACE + 1;
Do we still need this constant?

[Yang, 2016-08-09 13:34:35]

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
File src/snapshot/snapshot-common.cc (left):

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
src/snapshot/snapshot-common.cc:120: for (const auto& reservation :
context_snapshots->at(i)->Reservations()) {
Can we preserve the statistics here some how?

[Michael Lippautz, 2016-08-09 13:55:32]

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
File src/snapshot/snapshot-common.cc (left):

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
src/snapshot/snapshot-common.cc:120: for (const auto& reservation :
context_snapshots->at(i)->Reservations()) {
On 2016/08/09 13:34:34, Yang wrote:
> Can we preserve the statistics here some how?

Done.

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot.h
File src/snapshot/snapshot.h (right):

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
src/snapshot/snapshot.h:88: // [6] number of contexts N
On 2016/08/09 13:26:29, Yang wrote:
> Please update the indices here.

Done.

https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
src/snapshot/snapshot.h:97: static const int kNumPagedSpaces = LAST_PAGED_SPACE
- FIRST_PAGED_SPACE + 1;
On 2016/08/09 13:26:29, Yang wrote:
> Do we still need this constant?

Removed.

[Yang, 2016-08-10 05:30:58]

On 2016/08/09 13:55:32, Michael Lippautz wrote:
>
https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
> File src/snapshot/snapshot-common.cc (left):
> 
>
https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot-...
> src/snapshot/snapshot-common.cc:120: for (const auto& reservation :
> context_snapshots->at(i)->Reservations()) {
> On 2016/08/09 13:34:34, Yang wrote:
> > Can we preserve the statistics here some how?
> 
> Done.
> 
> https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot.h
> File src/snapshot/snapshot.h (right):
> 
>
https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
> src/snapshot/snapshot.h:88: // [6] number of contexts N
> On 2016/08/09 13:26:29, Yang wrote:
> > Please update the indices here.
> 
> Done.
> 
>
https://codereview.chromium.org/2013713003/diff/700001/src/snapshot/snapshot....
> src/snapshot/snapshot.h:97: static const int kNumPagedSpaces =
LAST_PAGED_SPACE
> - FIRST_PAGED_SPACE + 1;
> On 2016/08/09 13:26:29, Yang wrote:
> > Do we still need this constant?
> 
> Removed.

Thanks. LGTM.

[Michael Lippautz, 2016-08-10 08:59:40]

Patchset #23 (id:740001) has been deleted

[Michael Lippautz, 2016-08-10 09:00:10]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-10 09:00:21]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Michael Lippautz, 2016-08-10 09:35:04]

The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-10 09:35:14]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-10 09:58:44]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-10 09:58:47]

Dry run: This issue passed the CQ dry run.

[Hannes Payer (out of office), 2016-08-10 11:40:29]

lgtm

[Michael Lippautz, 2016-08-10 11:43:02]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
==========

[Michael Lippautz, 2016-08-10 11:43:36]

The CQ bit was checked by mlippautz@chromium.org

[Michael Lippautz, 2016-08-10 11:43:36]

The patchset sent to the CQ was uploaded after l-g-t-m from ulan@chromium.org,
yangguo@chromium.org
Link to the patchset: https://codereview.chromium.org/2013713003/#ps780001
(title: "Disable ReleaseOverReservedPages when no snapshot is available")

[commit-bot: I haz the power, 2016-08-10 11:43:48]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-10 11:45:27]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
==========

[commit-bot: I haz the power, 2016-08-10 11:45:28]

Committed patchset #24 (id:780001)

[commit-bot: I haz the power, 2016-08-10 11:45:38]

Description was changed from

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
==========

to

==========
[heap] Switch to 500k pages

- Decrease regular heap object size to 300k, keeping the same ration (60%)
  between this limit and page size.

In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.

Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)

BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org

Committed: https://crrev.com/ffe5c670e1559d11e7b252e15fec38765e7dbe4f
Cr-Commit-Position: refs/heads/master@{#38533}
==========

[commit-bot: I haz the power, 2016-08-10 11:45:40]

Patchset 24 (id:??) landed as
https://crrev.com/ffe5c670e1559d11e7b252e15fec38765e7dbe4f
Cr-Commit-Position: refs/heads/master@{#38533}

[Michael Lippautz, 2016-08-10 12:12:29]

A revert of this CL (patchset #24 id:780001) has been created in
https://codereview.chromium.org/2229403003/ by mlippautz@chromium.org.

The reason for reverting is: Failures on waterfall:

e.g.
http://build.chromium.org/p/client.v8/builders/V8%20Linux64%20TSAN/builds/11134.